Dicom Metadata Extract

NVIDIAOfficialCreated by NVIDIA MedTech Team

Skill
Tool integration
Science and research
Apache-2.0
Medium risk
Has scripts

From the author

Used for extracting selected metadata from one DICOM file and flagging standard-tag PHI presence. Not for anonymization or clinical use.

— the description by NVIDIA MedTech Team

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

The dicom-metadata-extract skill provides a programmatic interface for reading DICOM files to retrieve core metadata and identify the presence of Protected Health Information (PHI) based on a subset of standard tags. It is designed for technical evaluation and data processing workflows rather than clinical or regulatory use. The skill outputs structured JSON containing modality, transfer syntax, and grouped study information, while explicitly warning against its use for anonymization or clinical decision-making.

What it does

Extracts study, series, and image metadata from a single DICOM file.
Flags the presence of PHI based on standard DICOM tags.
Outputs metadata in a structured JSON format.
Provides a command-line interface for integration into automated pipelines.
Supports validation of output quality through dedicated verification tools.

When to use it

When you need to inspect DICOM file headers for technical metadata.
When performing initial data screening to identify standard PHI tags.
When building automated data processing pipelines for medical imaging research.
When verifying the structure of DICOM files during development or testing.

Advantages

Provides a lightweight and standardized way to access DICOM metadata.
Includes built-in PHI flagging for common standard tags.
Integrates easily with existing Python-based medical imaging workflows.
Supports trusted execution patterns for evidence-based review.

Usage examples

“Extract metadata from the file scan_01.dcm and save it to output.json.”: The agent runs the extraction script on the specified file and generates a JSON file containing the extracted metadata and PHI flags.
“Check if the DICOM file at /data/ct_scan.dcm contains any standard PHI tags.”: The agent processes the file and returns a report indicating whether PHI was detected and which specific tags were identified.

Install

Download the skill as a ZIP:
https://mp-staging.autogineer.com/agent-skills/nvidia/dicom-metadata-extract/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip dicom-metadata-extract.zip -d .agents/skills

Documentation by NVIDIA MedTech Team

DICOM Metadata Extract

Purpose

Used for extracting selected metadata from one DICOM file and flagging standard-tag PHI presence. Not for anonymization or clinical use.
Use the wrapper exactly as documented; do not replace the upstream entrypoint with a handwritten implementation.
Manifest I/O: inputs are dicom_path; outputs are metadata_json.

Instructions

Read skill_manifest.yaml before changing arguments, side effects, or validation gates.
Run scripts/extract_metadata.py through the documented command below; keep outputs under a caller-provided run directory.
If a host agent exposes run_script, use run_script("scripts/extract_metadata.py", args=[...]); otherwise run the Bash/Python command shown below.
Check the emitted JSON and run medagent.verifiers.dicom_metadata_quality_v1 on evidence packs before treating the run as reviewed evidence.

Available Scripts

Script	Purpose	Arguments
`scripts/extract_metadata.py`	Primary entrypoint declared by skill_manifest.yaml.	`PATH_TO_DICOM [--output OUT.json]`

Prerequisites

Runtime requirements: Python packages listed in runtime.side_effects.pip_packages.
Run commands from the repository root unless an existing section below says otherwise.

Limitations

Small PS3.15-inspired standard-tag subset only; not a complete Basic Application Confidentiality Profile implementation.
Private tags not checked
Burnt-in pixel PHI not detected
Multi-frame handling minimal
Not for clinical deployment, regulatory de-identification, autonomous diagnosis, patient-facing use.

Troubleshooting

Error	Cause	Fix
Missing dependency or import error	Runtime package drift from `skill_manifest.yaml`.	Install the packages declared in the manifest or use the documented setup command.
Empty or schema-invalid output	Wrong input path, unsupported modality, or upstream failure.	Re-run with a known fixture and inspect the wrapper JSON plus stderr.
Validation gate failure	Output violated a declared engineering invariant.	Keep the failed evidence pack and use the gate message to repair inputs or wrapper code.

Reads one DICOM file with pydicom and emits JSON on stdout.

python scripts/extract_metadata.py PATH_TO_DICOM
python scripts/extract_metadata.py PATH_TO_DICOM --output result.json

Output includes transfer_syntax, modality, grouped study/series/image metadata, phi_present, and phi_tags_found.

Use this as the smallest end-to-end example of a Medical AI Skills skill. Do not use it for anonymization, private-tag review, pixel PHI detection, or clinical interpretation.

For second-pass evidence review, generate a trusted run:

python -m eval_engine.run_trusted skills/dicom-metadata-extract \
  --fixture skills/dicom-metadata-extract/fixtures/sample_ct.dcm \
  --out runs/dicom_metadata_trusted

Files

AGENTS.md508 B
ATTRIBUTION.mdAdded by AutoGineer937 B
BENCHMARK.md7.0 KiB
LICENSELicence file added by AutoGineer11.2 KiB
SKILL.md3.2 KiB
evals/baseline.yaml2.2 KiB
evals/evals.json1.8 KiB
fixtures/generate_sample.pyScript3.0 KiB
scripts/extract_metadata.pyScript5.3 KiB
skill-card.md4.5 KiB
skill.oms.sig6.1 KiB
skill_manifest.yaml5.0 KiB
tests/test_basic.pyScript2.5 KiB
validators/output_schema.json2.3 KiB

BENCHMARK.md

Skill Benchmark: dicom-metadata-extract

✅ Overall verdict: PASS — Recommended for publication

Publication Recommendation

Recommended for publication based on the completed evaluation evidence in this report.

Evaluation Metadata

Skill: dicom-metadata-extract
Evaluation date: 2026-09-14
Evaluator version: 1.5.6
Agents: Claude Code (aws/anthropic/bedrock-claude-opus-4-8), Codex (openai/openai/gpt-5.5)
Tasks: 2 evaluation tasks (2 positive)
Dataset digest: sha256:1c4acd30852c7efd65d9d439bca3bcb67f590e7024f3e1cdfbbfc3f6660f1771 (skill-evaluator-dataset-snapshot/1)
Attempts per task: 3
Environment: k8s-sandbox
Tier 2 evidence: required for publication
Tier 3 evidence: required for publication

Each task attempt ran in its own isolated sandbox pod.

What This Report Answers

The three-tier evaluation checks whether the skill:

is safe to use;
produces correct answers;
is discovered and activated when needed;
helps the agent complete the user's goal and expected workflow; and
avoids wasted skill and tool usage.

Results at a Glance

Measure	Claude Code (Baseline → Skill Uplift)	Codex (Baseline → Skill Uplift)
Overall	72.3% — baseline ran, but no comparable score was available; uplift unavailable	77.0% — baseline ran, but no comparable score was available; uplift unavailable
Security	100.0% → 50.0% (-50.0 points)	100.0% → 50.0% (-50.0 points)
Correctness	25.0% → 80.0% (+55.0 points)	15.0% → 100.0% (+85.0 points)
Discoverability	91.5% — baseline ran, but no comparable score was available; uplift unavailable	75.0% — baseline ran, but no comparable score was available; uplift unavailable
Effectiveness	28.3% → 55.0% (+26.7 points)	28.3% → 85.0% (+56.7 points)
Efficiency	84.8% — baseline ran, but no comparable score was available; uplift unavailable	74.8% — baseline ran, but no comparable score was available; uplift unavailable

How to read this table: baseline is the same task attempted without the target skill. Scores are rounded to one decimal; threshold-adjacent values use additional precision so their displayed band matches the verdict. Uplift is derived from those displayed scores and shown in percentage points.

Example: 47.0% → 92.0% (+45.0 points) means the skill-assisted run scored 92.0%, 45.0 percentage points above its 47.0% no-skill baseline.

Token Usage

Actual Tier 3 execution usage is reported for every observed agent/case pair and both conditions.

Agent	Dataset case	With skill	Without skill	Delta	Change	Coverage
claude-code	All cases	647,052	605,045	N/A	N/A	skill 2/2; base 4/4
claude-code	extract-modality-and-phi-flag	212,638	482,358	N/A	N/A	skill 1/1; base 3/3
claude-code	no-anonymization-claim	434,414	122,687	+311,727	+254.08%	skill 1/1; base 1/1
codex	All cases	642,885	382,685	N/A	N/A	skill 2/2; base 4/4
codex	extract-modality-and-phi-flag	391,281	265,946	N/A	N/A	skill 1/1; base 3/3
codex	no-anonymization-claim	251,604	116,739	+134,865	+115.53%	skill 1/1; base 1/1
ALL AGENTS	Dataset aggregate	1,289,937	987,730	N/A	N/A	skill 4/4; base 8/8

Prompt tokens include cached reads, so total tokens are prompt + completion (cached is not added twice). The Efficiency score uses (prompt - cached) + completion. N/A means the relevant trajectory counters were not available; coverage is never estimated.

Tier Status

Tier	Purpose	Status	Evidence
Tier 1	Static validation	PASSED WITH OBSERVATIONS	11 validator(s); 16 finding(s)
Tier 2	Semantic deduplication	PASSED	2 validator(s); 0 finding(s)
Tier 3	Live agent evaluation	PASS	2 agent(s); 2 task(s)

Findings and Observations

<details> <summary>Show detailed findings and successful checks</summary>

MEDIUM PII/ip_addresses: Public IP address (fixtures/generate_sample.py:33)
MEDIUM PII/ip_addresses: Public IP address (fixtures/generate_sample.py:49)
MEDIUM PII/ip_addresses: Public IP address (fixtures/generate_sample.py:51)
MEDIUM SCHEMA/body_recommended_section: Missing recommended section: '## Examples' (skills/dicom-metadata-extract/SKILL.md)
MEDIUM SECURITY/Unknown (SQP-2): The skill reads DICOM files, extracts PHI-containing metadata (e.g., patient name, DOB, MRN — standard DICOM tags), and (SKILL.md:51)
11 additional finding(s) are available in the full evaluation artifacts.

</details>

Scoring Methodology

<details> <summary>Show dimension definitions, source signals, and thresholds</summary>

Dimension	Question	Scored signals
Security	Is it safe to use?	`security` (100%)
Correctness	Is the answer correct?	`accuracy` (100%)
Discoverability	Was the right skill loaded when needed?	`skill_execution` (100%)
Effectiveness	Did the skill help complete the task?	`goal_accuracy` (50%) + `behavior_check` (50%)
Efficiency	Did it avoid wasted tool calls and token usage?	`skill_efficiency` (50%) + `token_efficiency` (50%)

Dimension bands: PASS at 50% or above; NEUTRAL from 40% to below 50%; FAIL below 40%.
Overall Tier 3 lift: PASS at +5 points or more; FAIL at -10 points or less; values between those bands are NEUTRAL.
Overall verdict: PASS only when every configured dimension passes for at least one supported agent. Lift is reported as diagnostic evidence and does not override this gate.
The 50% attempt pass threshold is a separate per-task gate; it is not the dimension pass threshold.
Effectiveness is the equal-weight mean of goal completion (goal_accuracy) and expected workflow adherence (behavior_check).
Efficiency is 50% tool-call productivity (the backward-compatible skill_efficiency wire id) and 50% token_efficiency. Positive-case skill routing is scored under Discoverability, not Efficiency; a negative case without a routing target is N/A. N/A sources are omitted, remaining weights are renormalized, and the dimension is marked partial.

Signals present in this run:

security (Security): unsafe operations, secret leakage, and unauthorized access.
skill_execution (Skill Execution): whether the expected skill was selected, decoys were avoided, and the workflow executed.
skill_efficiency (Tool Productivity): tool-call productivity (legacy wire id; routing is scored under Discoverability).
accuracy (Accuracy): final-answer correctness against the reference answer.
goal_accuracy (Goal Accuracy): whether the user's goal was achieved.
behavior_check (Behavior Check): whether the expected workflow behavior was followed.
token_efficiency (Token Efficiency): actual uncached prompt plus completion usage (50% of Efficiency).

</details>

Freshness

Regenerate this benchmark when the skill, evaluation dataset, target agent/model, evaluator version, environment, or scoring policy changes.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Declares allowed tools, Needs Python

Security

What our scanner noticed· Medium risk

Large base64-encoded blobs
Scripts that your agent may run

This skill includes scripts that your agent may run. Read them before you use it.

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by

NVIDIA MedTech Team

Publisher

NVIDIA (Official)

Source

NVIDIA/skills · skills/dicom-metadata-extract · commit fce3cca

Licence

Apache License 2.0 (redistribution allowed)

LICENSE (added by AutoGineer from LICENSE-APACHE)
ATTRIBUTION.md (written by AutoGineer)

Synced

2026-11-08

AutoGineer is not affiliated with or endorsed by NVIDIA MedTech Team or NVIDIA.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download dicom-metadata-extract.zip